In condition-based monitoring, sensor readings are commonly used to determine the wear of vehicular components. Therefore, such readings often serve as predictors of wear related component failures. However, with respect to non-wear related component failures, such predictors are lacking.
Accordingly, a need exists for techniques for generating estimates of non-wear related failure risk for a vehicular component based on input sets of time series.